A N N A L S O F T H E N E W Y O R K A C A D E M Y O F S C I E N C E S
Issue: The Esophagiome
The esophageal microbiota in health and disease
Vincenzo Di Pilato,1Giancarlo Freschi,1,2Maria Novella Ringressi,1,2Lucia Pallecchi,3 Gian Maria Rossolini,3,4,5,6and Paolo Bechi1,21Department of Surgery and Translational Medicine, University of Florence, Florence, Italy.2Gastrointestinal Surgery Unit, Florence Careggi University Hospital, Florence, Italy.3Department of Medical Biotechnologies, University of Siena, Siena, Italy.4Department of Experimental and Clinical Medicine, University of Florence, Florence, Italy.5Microbiology and Virology Unit, Florence Careggi University Hospital, Florence, Italy.6Don Carlo Gnocchi Foundation, Florence, Italy
Address for correspondence: Vincenzo Di Pilato, Department of Surgery and Translational Medicine, University of Florence, Largo Brambilla No. 3, 50134 Florence, Italy. vdipilato@yahoo.it
The esophageal mucosa is among the sites colonized by human microbiota, the complex microbial ecosystem that colonizes various body surfaces and is increasingly recognized to play roles in several physiological and pathological processes. Our understanding of the composition of the esophageal microbiota in health and disease is challenged by the need for invasive sampling procedures and by the dynamic nature of the esophageal environment and remains limited in comparison with the information available for other body sites. Members of the genus Streptococcus appear to be the major components of the microbiota of the healthy esophagus, although the presence of several other taxa has also been reported. Dysbiosis, consisting of enrichment in some Gram-negative taxa (including Veillonella, Prevotella, Haemophilus, Neisseria, Campylobacter, and Fusobacterium), has been reported in association with gastroesophageal reflux disease and is hypothesized to contribute to the evolution of this condition toward Barrett’s esophagus (which is the most common esophageal precancerous lesion) and, eventually, adenocarcinoma. Some Campylobacter species (mostly C. concisus) are also putatively involved in the progression of disease toward adenocarcinoma. However, variable findings have recently been reported in additional studies. Causative relationships between dysbiosis or specific bacterial species and esophageal diseases remain controversial and warrant further investigations.
Keywords: esophageal diseases; microbiota; esophagitis; Barrett’s esophagus; esophageal adenocarcinoma; esophageal
squamous cell carcinoma
Introduction
The human body has symbiotic relationships with a complex and diverse array of microbial commu-nities at various nonsterile body sites, which overall constitute the human microbiota (HM). The sites colonized by the HM include skin and various mucous membranes (e.g., mouth, upper respiratory tract, lower genitourinary tract, gastrointestinal tract). The composition of the HM exhibits a remarkable variability at various sites, among differ-ent individuals, and at differdiffer-ent times, depending on several factors that are only partially understood.1–3 Until recently, knowledge of the HM com-position was largely hampered by the limited resolution allowed by cultural methods, which cannot recover a large number of components of the
HM. The advent of culture-independent metage-nomic methods capable of also characterizing the nonculturable fraction of complex microbial com-munities, in combination with high-throughput next-generation sequencing (NGS) technologies, has provided a fundamental breakthrough in studying the composition of the HM.4–6
The HM is now considered an additional appara-tus of the human body, capable of modulating sev-eral developmental, metabolic, and immunological pathways and playing a fundamental role in defense against colonization by microbial pathogens.1,7–10 It is also clear that HM alterations (a condition termed dysbiosis) can be associated with disease.11 The most paradigmatic example is represented by the association between dysbiosis of the intestinal microbiota, promoted by antibiotic treatments, and
doi: 10.1111/nyas.13127
21
Ann. N.Y. Acad. Sci. 1381 (2016) 21–33C2016 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences.
Clostridium difficile infection (CDI),12–14 which is the leading cause of diarrhea among hospitalized patients and has become a major public health issue in several settings.15 Under these circum-stances, the conventional treatment of CDI, based on antibiotics active against C. difficile, may sus-tain a vicious cycle wherein the dysbiotic status per-sists and favors recurrences of CDI. In this context, reconstitution of the intestinal microbiota by fecal microbiota transplantation has proved extremely successful and has definitively confirmed the role of dysbiosis in the pathogenesis of CDI.16–18 An association with altered composition of the HM has now also been reported for several other dis-eases, including inflammatory bowel disdis-eases,19 col-orectal cancer,20obesity,21diabetes,22nonalcoholic steatohepatitis,23 bacterial vaginosis,24 atopic der-matitis, and psoriasis,9,25 although in these cases a causative relationship remains to be definitively established.
The esophageal mucosa is one of the sites colo-nized by HM, and consequently there is an interest in understanding a possible role of HM alterations in esophageal diseases. Some articles have recently reviewed various aspects of this topic.26–31Here, we provide an updated review of current knowledge on the esophageal microbiota in health and in different esophageal diseases and discuss the major challenges encountered in investigation of this topic.
The healthy esophagus and its microbiota
Structure and physiology of the healthy esophagus
The esophagus is a muscular conduit of approxi-mately 20–27 cm in length, with an internal mucosa carrying a stratified squamous epithelial layer. It plays the fundamental role of transferring the ali-mentary bolus from the pharynx to the stomach. At its distal end, a complex valve mechanism allows passage of the bolus into the stomach and restricts reflux of the gastric contents back into the esopha-gus. The esophagus has a virtual narrow cavity that is empty and collapsed, except during swallowing. The epithelial layer of the esophageal mucosa is nor-mally wet with saliva for the whole length and, as a result, the pH is usually around 7. However, reflux of gastric material may occur and cause a sudden lowering of pH values (down to 2). This is mostly limited to the distal third of the organ, and, nor-mally, acidification is rapidly buffered by esophageal
motility. The potentially harmful effect of reflux on the esophageal mucosa depends on the composi-tion of the refluxate (commonly gastric acid and ingesta, less frequently bile and pancreatic enzymes) and the time of contact. In addition to mucosal damage, reflux is capable of inducing micromental variations. Therefore, whereas the environ-ment of the proximal esophageal mucosa is similar to the oral one, an environment that is intermediate between the oral-like one of the proximal esophagus and the gastric one has to be expected in the distal esophageal mucosa.
The microbiota of the healthy esophagus The esophageal mucosa hosts a resident micro-biota, although in a smaller population as compared with other districts of the gastrointestinal tract. In fact, the HM of the gastrointestinal tract exhibits substantial qualitative and quantitative differences, with populations ranging from 10 cells per g/mL of sampled material in the esophagus and stomach to 1012 per g/mL of sampled material in the large intestine.32,33
Considering the anatomic structure and function, it was originally unclear whether the esophagus was associated with a defined microbiota. The first stud-ies on the esophageal microbiota, dating back to the early 1980s and based on cultural methods, demon-strated that the esophagus was not a sterile site and did not simply contain a transient microbial popu-lation originating from the oral cavity by swallow-ing or from the stomach by gastroesophageal reflux (GER).34–36 Consistently, it was later observed that bacteria were closely associated with the esophageal mucosal surface, confirming the presence of a resi-dent microbiota at this site.37More recently, knowl-edge regarding the composition of the microbiota of healthy individuals has been expanded through the use of investigations based on metagenomic approaches.4–6However, the difficulties in obtain-ing samples at this body site, compared with other sites, have limited the number of studies on the esophageal microbiota.
Few studies have focused on the composition of the microbiota of the normal esophagus alone (Table 1),37–40but additional information has come from studies investigating the esophageal micro-biota in disease and including healthy patients as controls for comparison (Tables 26,41–47and 348–50). The available information on the microbiota of the
Table 1. Studies reporting the characterization of esophageal microbiota in the healthy individuals Year No. of studied subjects Exclusion
criteria Method Samples
Sampled esophageal
tract Main findings Reference
1998 20 Esophageal abnormalities Culture-based Aspirate Middle third Streptococcus, Enterococcus, Staphylococcus, Klebsiella
38 2004 4 Recent antibiotic use
Previous gastric/ esophageal surgery Active oral cavity infection
Culture-independent (sequencing of 16S rRNA gene clones, 25–30 clones/ sample analyzed)
Biopsy 2 cm above the squamocolumnar junction Streptococcus, Prevotella, Veillonella 37 2012 15 Esophageal abnormalities Increased risk for endoscopic
complications Culture-independent (NGS, V2–V3 region of 16S rRNA genes) Biopsy and esophageal string test
Mid to distal Streptococcus, Prevotella, Veillonella
39
2013 40 Age<20 or >75 years
Recent use of antibiotics or PPIs
Gastroesophageal diseases Ongoing infections
Culture-based Biopsy and brush
5 cm below the upper esophageal sphincter, 2 cm above the squamocolumnar junction Streptococcus, Neisseria, Haemophilus, Prevotella 40
NGS, next-generation sequencing; PPIs, proton pump inhibitors.
healthy esophageal mucosa, in terms of prevalence and proportions of bacterial taxa, is summarized in Figure 1 and briefly discussed below, with the caveat that different methodological approaches, differences in the sampled areas within the esoph-agus, and the heterogeneity of inclusion/exclusion criteria used in the various studies do not allow fine-tuned comparisons and make it difficult to reach a consensus on the comprehensive microbiota composition of the healthy esophagus.
The presence of Streptococcus spp. was constantly reported, and in high proportions, by all studies: therefore, members of this genus appear to be a dominant taxon in the microbiota of the normal esophagus. Other bacterial genera frequently detected (at least in half of the studies) in associ-ation with streptococci, although usually in lower proportions, include Prevotella, Fusobacterium, and Veillonella. The presence of other genera (e.g., Haemophilus, Neisseria, Gemella, Granulicatella, Lactobacillus, Actinomyces, Staphylococcus, Bac-teroides, and Porphyromonas) was less frequently reported, and some genera were only reported by single studies (Fig. 1).
Notably, the prevalence of Streptococcus, Fusobac-terium, Veillonella, and Prevotella has consis-tently been reported in several studies based on either culture-dependent or culture-independent approaches and also on different specimens (biop-sies, brushes, aspirates) (Tables 1–3), thus provid-ing a strong indication of their contribution to the composition of the core microbiota that colonizes the healthy esophageal mucosa. The dominance of streptococci and the frequent presence of other taxa
typical of the oropharyngeal microbiota have been related to the composition of the microbial com-munities of the oropharyngeal cavity, where a high prevalence of streptococci is found, together with Veillonella, Fusobacterium, Gemella, Granulicatella, and Rothia,51,52and have supported the notion that the esophageal microbiota is primarily of oral ori-gin. However, not all oral bacteria are apparently able to colonize the esophageal mucosa, while sev-eral members of the esophageal microbiota are not present or are underrepresented in the oral cavity, pointing to a diverse microbiota adaptation in the two body sites.37,39,40,49
Major esophageal diseases
The environment of the esophagus and its patho-logic states are closely connected. There is no doubt that, apart from esophageal squamous cell can-cer (ESCC), whose pathogenesis remains largely unknown (although related to cigarette smoking and alcohol consumption), all the most important esophageal diseases are associated with GER and GER-related environmental changes. These are rep-resented by gastroesophageal reflux disease (GERD) with or without esophagitis, Barrett’s esophagus (BE), and esophageal adenocarcinoma (EAC), with GERD generally representing the first step in the evolution of disease toward BE and EAC.
The prevalence of GERD has been rapidly increas-ing in Western countries, whereࣈ20% of adults today may be affected by this disease.53,54The pres-ence of short-duration reflux episodes is considered physiologic, but GERD occurs when reflux becomes symptomatic and/or causes esophageal mucosal
Table 2. Studies reporting the characterization of esophageal microbiota in major esophageal diseases (E, BE, EAC, ESCC) Year No. of studied subjects Disease group (No.) Control group (No.) Exclusion
criteria Method Samples
Sampled esophageal
tract Main findings Reference
1981 79 EC (79) – – Culture-based Aspirate – Detected bacteria included Bacteroidetes (39%), Streptococcus (10%), and coliforms (8%) 34 1982 12 EAC (7) ESCC (5)
– – Culture-based Biopsy Proximal resection line after cancer excision
Streptococcus and Bacteroidetes most
prevalent taxa
35
1983 101 EC (50) NE (51) Use of antibiotics Culture-based Aspirate – No differences in the presence/absence of specific taxa in EC vs. controls.
Streptococcus and Peptococcus the most
prevalent taxa
36
2004 20 EC (20) – – Culture independent (sequencing of 16S rRNA gene clones of most members of the oral microbiota, except Neisseria,>10
clones/sample analyzed)
Biopsy Cancerous and noncancerous tissues from each patient
Treponema denticola, Streptococcus mitis, and Streptococcus anginosus prevalent in both cancerous and noncancerous tissues and suggested to have roles in carcinogenic process
68
2005 24 E (12) BE (3)
NE (9) Recent antibiotic use Previous gastric/
esophageal surgery Active oral cavity
infection
Culture independent (sequencing of 16S rRNA gene clones, two clones/sample analyzed)
Biopsy 2 cm above the squamocolumnar junction (NE and E), 2 cm above the gastroesophageal junction (BE) Proof-of-concept study demonstrating the presence of a nontransient microbiota in the reflux-exposed esophageal mucosa and the importance of culture-independent approaches 41 2007 14 BE (7) NE (6) E (1)
Recent antibiotic use Culture-based Biopsy and aspirate
Middle to distal High-level colonization by Campylobacter (mostly C. concisus) in BE, absence in controls 42 2009 34 E (12) BE (10)
NE (12) Recent antibiotic use Previous gastric/
esophageal surgery Active oral cavity
infection HIV infection
Culture-independent (sequencing of 16S rRNA gene clones, 200 clones/sample analyzed)
Biopsy 2 cm above the squamocolumnar junction Type I microbiota (dominated by Gram-positive taxa of the phylum Firmicutes) in normal esophagus, type II microbiota (enriched in Gram-negative taxa) in E and BE 43 2013 18 E (6) BE (6) NE (6) Recent use of antibiotics or PPIs Gastric/esophageal surgery Active oral cavity
infection HIV, HBV, or HCV
infection
Culture-independent (sequencing of 16S rRNA gene clones, about 24 clones/sample analyzed)
Biopsy 1 cm above the gastroesophageal junction Enrichment of Prevotella, Fusobacterium, Veillonella, and Neisseria in BE 44 2013 34 E (8) BE (8) EAC (10) NE (8) Recent use of antibiotics or PPIs Culture-based plus qRT-PCR for detection of specific taxa in a larger cohort
Biopsy 5 cm above the esophagogastric junction, or at the upper limit of BE or at site of pathology Campylobacter (mostly C. concisus) expansion in E and BE compared to AD and controls (suggested to be specific to reflux-exposed mucosa) 45 2014 34 E (13) BE (6) NE in GERD (15) Recent use of antibiotics or PPIs Culture-independent (NGS, V6–V7 region of 16S rRNA genes) Biopsy Normal-appearing mucosa above the inflammation site or BE No significant difference between microbiota of normal and abnormal esophagus. Relevant effect of PPI treatment 46 Continued
Table 2. Continued Year No. of studied subjects Disease group (No.) Control group (No.) Exclusion
criteria Method Samples
Sampled esophageal
tract Main findings Reference
2015 12 BE (12) – EAC Culture-independent (NGS, V3–V4 region of 16S rRNA genes)
Biopsy and brush Normal-appearing and BE mucosa Streptococcus and Prevotella dominant species. No substantial intraindividual difference between squamous and BE mucosal microbiota 59 2016 26a BE (13) EAC (5) NE in GERD (8) – Culture-independent (PCR–ESI–MS-TOF)
Biopsy In some cases, both normal and diseased mucosa
Higher prevalence of
Streptococcus pneumoniae in
GERD and BE, compared to EAC
47
–, absent or not specified.
aThe authors reported 28 patients, which possibly included two patients with dysplasia, in addition to those indicated in the table. E, reflux-related esophagitis; EC, esophageal carcinoma (not further specified); EAC, esophageal adenocarcinoma; ESCC, esophageal squamous cell carcinoma; GERD, gastroesophageal reflux disease; NE, normal esophagus; BE, Barrett’s esophagus; NGS, next-generation sequencing; PCR–ESI–MS-TOF, PCR-based electron spray ionization–time-of-flight mass spectrometry; PPIs, proton pump inhibitors; qRT-PCR, quantitative real-time PCR.
damage associated with inflammation (esophagi-tis). This seems to be attributable to an increased GER/mucosa contact time, which overwhelms the protective mechanisms of the esophageal mucosa, including esophageal clearing, salivation, bicarbon-ate secretion by esophageal submucosal glands, and the histologic/functional endothelial config-uration. Salivation may play a major role in the protective effect of esophageal mucosa.55 In fact, it does not appear incidental that esophagi-tis, BE, and EAC are generally limited to the lower third of the esophagus, where the saliva-tion effect is lower and the environment is inter-mediate between the esophageal and the gastric environments.
Although the role of the individual components of GER has been debated,56there is no doubt that chronic exposure to reflux may determine the onset of esophagitis and its progression, although not frequently, to BE and EAC. BE is represented by the replacement of the stratified squamous epithelium of the distal esophagus with an intestinal columnar metaplastic epithelium. It is the most common esophageal precancerous lesion and an intermediate step in the sequence GER–esophagitis–intestinal metaplasia–dysplasia–EAC. EAC, therefore, like colon cancer, identifies a multistep cancer model that is being thoroughly studied. The evolution of the sequence has been shown to largely depend upon inflammation and its mediators, such as cyclooxygenase-2 (COX-2).57 However, although mucosal inflammation can be directly caused
by GER, it could also be caused by esophageal microbiome variations induced by GER (see below).
The microbiota in diseased esophagus and its potential role in diseases
Microbiota in GERD, BE, and esophageal cancers
As previously mentioned, GERD-related esophagi-tis and Barrett’s metaplasia are the major conditions associated with chronic reflux of acidic material from the stomach into the lower part of the esoph-agus. Under these circumstances, a modification of the esophageal microbiota is expected to occur owing to the altered environment, and it has been hypothesized that the modification of the micro-biota could contribute to disease persistence and/or progression.26–31The relevant literature available on this subject is summarized in Table 2 and discussed below.
A positive association between the load of bac-terial colonization and severity was first reported in 2004 by a study based on Gram staining of esophageal biopsies in reflux-related esophagitis and BE.58Shortly thereafter, Pei et al.41demonstrated the presence of a nontransient microbial population in GER-exposed esophageal mucosa and, at the same time, established the importance and feasibility of culture-independent approaches for investigating the complexity of the esophageal microbiota. A subsequent culture-based characterization of microbial communities in patients affected by BE compared to controls was carried out by Macfarlane
Table 3. Studies reporting the characterization of esophageal microbiota associated with some uncommon esophageal diseases Year No. of studied subjects Disease group (No.) Control group (No.) Exclusion
criteria Method Samples
Sampled esophageal
tract Main findings Reference
2002 25 CME (15) NE (10) Grade IV megaesophagus Recent antibiotic use GERD (for controls)
Culture-based Aspirate Distal Higher bacterial load and greater variability in CME (e.g., enrichment in Veillonella) compared to controls 48 2015 68 EoE (35) NE and mild non-EoE (33)
Use of antibiotics or oral steroids Immune or other inflammatory GI disorders Culture-independent (NGS, V1–V2 region of 16S rRNA genes)
Biopsy Distal Enrichment in Proteobacteria (Neisseria and Corynebacterium) and decrease of Firmicutes (Streptococcus and Atopobium) in EoE 49 2015 70 EoE (37) E (8) NE (25) Age<7 years Esophageal abnormalities Antibiotic use Increased risk for endoscopic
complications Culture-independent (NGS, V1–V2 region of 16S rRNA genes) Esophageal string test
Middle to distal Enrichment in Proteobacteria (Haemophilus) and decrease of Firmicutes in active EoE 50
CME, chagasic megaesophagus; E, reflux-related esophagitis; EoE, eosinophilic esophagitis; GI, gastrointestinal; NE, normal esoph-agus; NGS, next-generation sequencing; GERD, gastroesophageal reflux disease.
et al.42 In that study, although microbial growth was observed only from a small subset of samples (i.e., four and three samples from BE and controls, respectively), the authors observed a uniquely high level of colonization by Campylobacter spp. (namely C. concisus and C. rectus) in BE and hypothesized a potential role of this pathogen in disease progression to EAC.
In 2009, following a large-scale, culture-independent study (i.e., 6800 16S bacterial rRNA gene clones sequenced from 34 individuals), Yang et al.43 first identified two distinct patterns of esophageal microbiota associated with normal and abnormal esophageal mucosa. In particular, a microbiota dominated by Gram-positive taxa of the Streptococcus genus (phylum Firmicutes), named type I microbiota, was found to characterize the normal esophageal mucosa, while a shift toward a microbiota enriched in Gram-negative taxa (e.g., Veillonella, phylum Firmicutes; Prevotella, phylum Bacteroidetes; Haemophilus, Neisseria, and Campy-lobacter, phylum Proteobacteria; Fusobacterium, phylum Fusobacteria), named type II microbiota, was found to be associated with the esophageal mucosa in cases of GERD-related esophagitis and BE. The presence of distinct microbiota compositions in normal esophagus, GERD-related esophagitis, and BE was also confirmed in a sub-sequent study, performed with a similar approach, although with a lower discrimination power (i.e.,
424 bacterial 16S rRNA gene clones sequenced from 18 individuals).44 Indeed, in that study, Liu et al. observed that, compared with healthy con-trols, the esophageal microbiota in GERD-related esophagitis and BE was enriched in Prevotella (phylum Bacteroidetes) and Fusobacterium (phy-lum Fusobacteria), and showed differences in the relative abundance of Firmicutes (a decrease of Streptococcus and an enrichment in Veillonella, mostly in BE) and of Proteobacteria (an overall decrease, but with enrichment in Neisseria).44
On the other hand, the first study that used NGS technologies for characterization of the esophageal microbiota failed to detect significant differences between patients with normal mucosa and patients with reflux-related esophagitis or BE.46 These authors hypothesized that the lack of identification of the previously described type I and type II esophageal microbiota in healthy and diseased esophagus could be due to issues related to the experimental design and the methodological approach (e.g., in this study, all controls had GERD symptoms, and samples were obtained from normal-appearing mucosa above the diseased tissue,46while in the study by Yang et al.,43only one-third of controls had GERD symptoms, and samples were obtained from diseased mucosa). Moreover, this study first demonstrated a remarkable effect of proton pump inhibitor (PPI) treatment on the microbiota composition in the esophageal
Figure 1. Overview of the frequency and relative abundance of bacterial taxa (at the genus level) associated with the healthy esophageal mucosa reported by studies investigating the composition of the esophageal microbiota in healthy individuals. The presence of each bacterial genus is indicated by a circle whose size and color indicate the relative abundance: black-filled larger circles represent bacterial genera with a relative abundanceࣙ20%, gray-filled intermediate circles represent genera with a relative abundanceࣙ10% and <20%, and white-filled smaller circles represent genera with a relative abundance ࣙ1% and <10%. For Ref. 49, the reported frequencies were calculated on rarefied OTUs provided as supplemental material. Bacterial genera with an abundance<1% and cases where classification at the genus level was not available (e.g., the unidentified oral bacteria in Ref. 41) are not reported. Clustering of genera in different phyla are also shown. For Ref. 46, data at the genus level were not available, and the predominant phyla are indicated by striped boxes. When data for different types of esophageal specimens were available (e.g., in Refs. 39, 42, and 40), only those related to biopsy specimens were considered and included for comparison. For Ref. 40, the reported data refer to lower esophageal samples.
mucosa. In particular, PPI treatment was found to be associated with an enrichment in several members of the phylum Firmicutes (e.g., Clostridi-aceae and LachnospirClostridi-aceae) and with a decrease in members of the phylum Proteobacteria (e.g., Comamonadaceae). Since, in the study by Yang et al.,43 the information on PPI treatment was not available, it is possible that some of the observed shifts might be associated with PPI treatment, and this could represent an additional factor contribut-ing to the lack of consistency in the results of the two studies.
More recently, Gall et al.59 observed that Strep-tococcus and Prevotella dominated the esophageal microbiota of a cohort of BE patients, with no sub-stantial intraindividual differences between normal and metaplastic esophageal mucosa. In that study, the authors also pointed out that the Streptococcus to Prevotella ratio was significantly associated with some important risk factors for BE and EAC (e.g., waist-to-hip ratio, hiatal hernia length) and encour-aged further studies to investigate the clinical signif-icance of this finding.
The microbiota composition in esophageal can-cers has been investigated since the early 1980s,34 and during the last three decades it has been stud-ied using diverse dependent and culture-independent approaches (Table 2). The interest in such studies arises from the recognition that gas-trointestinal tract dysbiosis may play a role in dis-ease initiation and perpetuation.20,26–31,60–62 The mechanisms accounting for dysbiosis-mediated car-cinogenesis could involve both the induction of a cancer-promoting inflammatory response (mainly mediated by innate immunity) and the produc-tion of procarcinogenic compounds by the bac-terial taxa enriched in dysbiosis.20,26,28–31,60,61,63 In fact, innate immunity receptors, such as Toll-like receptors and Nod-Toll-like receptors, are major effectors in the maintenance of the gastrointesti-nal microbiota equilibrium, and their activation in the presence of dysbiosis has been demon-strated to trigger an inflammatory response pro-moting carcinogenesis.20,26,28–31,60,61Inflammation, in turn, is known to contribute to dysbiosis by several still incompletely understood mechanisms (e.g., inflammation-derived increased availability of nitrates may promote enrichment of facultative anaerobes, such as Enterobacteriaceae, which are able to use nitrates as electron acceptors), giving
rise to the vicious cycle of dysbiosis–inflammation– dysbiosis.20,26,28–31,60,61 Besides the activation of innate immunity and the possible direct proinflam-matory potential, certain bacteria are also capable of modulating carcinogenesis through the production of genotoxic compounds (e.g., cytolethal distend-ing toxin and colibactin, produced by members of Enterobacteriaceae, Campylobacter spp., and Heli-cobacter spp.)8,64,65 or other procarcinogenic metabolites (e.g., superoxide and hydrogen sul-fide, produced by Enterococcus faecalis and Fusobac-terium, respectively).66,67
The earlier studies on the correlation between microbiota and esophageal cancers suffered from the intrinsic limitations of culture-dependent approaches, the absence of control groups, and/or the lack of histological differentiation between ESCC and EAC, which are known to have dis-tinct risk profiles (Table 2).34–36 In 2004, Narikiyo et al.68 first used a culture-independent approach for the characterization of the microbiota in esophageal cancers, even though it is not clear if the study included ESCC, EAC, or both. In particular, Narikiyo et al. characterized the microbiota of nor-mal and cancerous esophageal tissue in 20 patients undergoing surgery for esophageal carcinoma and observed that both were consistently dominated by Treponema denticola and streptococci, including S. mitis and S. anginosus. On the basis of the detec-tion of those three species in a number of addi-tional biopsies of esophageal cancer from different countries (by polymerase chain reaction amplifica-tion with species-specific primers), and considering their proinflammatory potential, the authors sug-gested that these three species could play roles in the carcinogenic progress.
More recently, Blackett et al.,45 using a mixed culture-dependent and culture-independent app-roach, investigated the role of microbiota in the progression of EAC through the GER–esophagitis– intestinal metaplasia–dysplasia–adenocarcinoma cascade and observed a significant enrichment in Campylobacter (mostly C. concisus) in GERD and BE, compared with controls and esophageal EAC. On the basis of the observation that Campylobacter spp. (mostly C. concisus) are among the Gram-negative species consistently enriched in type II esophageal microbiota42,43,45 and on their overall higher prevalence in early stages of the EAC cascade (i.e., higher prevalence in GERD–esophagitis and
BE compared to EAC and controls),45 it has been proposed that this pathogen could play a role in EAC progression similar to that of Helicobacter pylori in gastric cancer.69 C. concisus can colonize the oral cavity and has recently been recognized as a poten-tial human pathogen, with diverse pathotypes being identified in addition to commensal strains.70,71 Chronic esophageal colonization by virulent C. concisus strains, favored by the inflammation and dysbiosis associated with GER, could induce augmentation of the inflammatory and metaplastic responses, favoring progression to EAC. Indeed, vir-ulent C. concisus strains have been shown to induce the expression of proinflammatory cytokines asso-ciated with carcinogenesis.69On the other hand, the lower prevalence of this pathogen in EAC would be consistent with the “driver–passenger” theory (i.e., a “driver” pathogen would be responsible for the initial steps in carcinogenesis, in turn favoring the proliferation of “passenger” bacteria with a com-petitive advantage in the tumor microenvironment and involved in disease progression)72 and would mirror the lower prevalence of H. pylori in gastric cancers.69
Besides the specific role of Campylobacter spp., the GERD-related dysbiosis, characterized by the shift from type I to type II esophageal micro-biota, has been suggested to contribute to EAC development.26,28–31 In the proposed pathogenetic scheme, a refluxated esophageal mucosa would ini-tially favor the switch between a Gram positive– dominated microbiota (i.e., type I microbiota) to a Gram negative–enriched one (i.e., type II microbiota), owing to the lower susceptibility of many Gram-negative species to acidic pH and bile salts. In turn, the increased abundance of Gram-negative bacteria would induce a strong inflamma-tory response through lipopolysaccharide (LPS, a component of Gram-negative outer membranes)-mediated activation of innate immunity, initiat-ing the vicious cycle of dysbiosis–inflammation– dysbiosis.26,28–31,60,61Exacerbated innate immunity signaling has been demonstrated for GERD— esophagitis, BE, and EAC,26,28–31and also in a recent animal model of EAC.47In addition, LPS has been associated with a dose-dependent decrease of the basal tone of the lower esophageal sphincter and with a delay in gastric emptying, both factors con-tributing to the maintenance and exacerbation of GER.26,28–31
Concerning ESCC, a negative correlation between esophageal microbial richness and esophageal squa-mous dysplasia (ESD) has recently been observed using a culture-independent approach,73suggesting that individuals with a lower esophageal microbial complexity could be more prone to develop ESD. An additional study performed using an NGS approach demonstrated an enrichment in Clostridiales and Erysipelotrichales (phylum Firmicutes) in the gastric corpus microbiota of patients affected by ESD and ESCC compared with controls, pointing to a possi-ble involvement of gastric dysbiosis in ESD–ESCC progression.74 Recently, a possible role for specific bacteria in the pathogenesis of ESCC has also been investigated by Gao et al.,75 who focused on the possible relationships between Porphyromonas gin-givalis, an oral pathogen thought to be involved in oral squamous cell carcinoma tumorigenesis, and ESCC, given the histological similarity of the two cancers. The authors reported that P. gingivalis could selectively infect the cancerous and adjacent esophageal mucosa of ESCC subjects but not the healthy mucosa of controls. Moreover, the presence of P. gingivalis was positively correlated with ESCC severity (i.e., cancer cell differentiation, metastasis) and poor prognosis (i.e., ESCC survival rate) and, therefore, it has been suggested that P. gingivalis may serve as biomarker of ESCC.
Microbiota in uncommon esophageal diseases
Two recent studies (Table 3)49,50investigated, using NGS approaches, the esophageal microbiota in eosinophilic esophagitis (EoE), an allergic disease characterized by a dense eosinophilic infiltrate within the esophageal mucosa.76Despite differences in the study design and experimental procedures, both studies identified a distinct esophageal micro-biota in active EoE (i.e., withࣙ15 eosinophils/high-power field) compared to controls.
Benitez et al.49 observed a shift in the relative abundance of specific taxa in active EoE com-pared with controls. In particular, Neisseria and Corynebacterium were significantly more abundant in the active EoE cohort, while Firmicutes (includ-ing Streptococcus and Atopobium) were consistently enriched among controls. Interestingly, no signif-icant differences were observed between inactive EoE (i.e.,< 15 eosinophils/high-power field) and controls, suggesting a major role of the esophageal
inflammatory state in modulating the local micro-biota composition. In support of this finding, the authors also observed an increase in the relative abundance of Granulicatella (phylum Firmicutes) and Campylobacter (phylum Proteobacteria) fol-lowing the reintroduction of highly allergenic foods in EoE patients previously treated with an empiric six food–elimination diet, a situation that leads to an increased esophageal inflammatory state. Overall concordant results, in terms of inflammation-mediated modulation of microbiota composition, were obtained in the study by Harris et al.,50 who observed an enrichment in Proteobacteria (mostly Haemophilus) and a decrease in Firmicutes in patients with active EoE compared with inactive EoE and controls. In the same study, a modulatory effect of PPI treatment was also observed in a small number of GERD patients (i.e., decrease of Firmi-cutes and increase in Proteobacteria), which did not confirm what was previously observed by Amir et al.46and points out the need to further investigate the effects of PPIs on the esophageal microbiota.
A single study has investigated the esophageal microbiota composition in patients affected by chagasic megaesophagus (a sequela of Trypanosoma cruzi infection), using a culture-dependent app-roach (Table 3).48In addition to a higher bacterial load related to chronic stasis, a greater variability and an apparent enrichment in Veillonella were observed, compared with controls. Acknowledging that chagasic megaesophagus represents an impor-tant risk factor for ESCC, the authors encouraged further studies to assess the potential role of a procarcinogenic dysbiosis in disease progression.
Conclusions and perspectives
An increasing number of studies published dur-ing the last decades have significantly advanced our knowledge on the composition of esophageal micro-biota in healthy subjects (Table 1) and have under-scored a possible role of dysbiotic conditions in the pathogenesis of some esophageal diseases, especially those related to GER (Tables 2 and 3). However, experimental findings have been variable, and a cur-rent understanding of these aspects remains limited. Therefore, additional studies are warranted to con-firm causative relationships between dysbiosis or specific bacterial species and GERD-related or other esophageal diseases and to establish the underlying molecular mechanisms.
When planning similar studies, it must be consid-ered that investigation of the esophageal microbiota in health and disease poses a number of method-ological and technmethod-ological challenges. First, obtain-ing samples requires invasive procedures that are not always acceptable, and taking longitudinal samples from the same subject can be difficult or impossible. These aspects have negatively affected the perfor-mance of long-term longitudinal studies and lim-ited the recruitment of large human cohorts, as demonstrated by the limited number of enrolled subjects in various studies (Tables 1–3). Second, the interactions within the esophageal environment are difficult to reproduce in vitro, which limits the possibility of studying the esophageal ecosystem using in vitro models,77–79 while the available ani-mal models to study esophageal diseases (e.g., the BE animal model based on a surgical anastomo-sis in the rat) are complex and have several short-comings, including species differences, the require-ment of surgical expertise, duration, and costs.80 Third, culture-dependent methods can contribute only partially to depicting the population structure of the resident microbiota, since several taxa are difficult or impossible to cultivate. In this context, the use of culture-independent approaches, based on molecular analysis of metagenomic libraries, has greatly helped in overcoming the limitations of con-ventional culture-dependent approaches, allowing for a higher analytical resolution and providing for a more comprehensive picture of the micro-bial communities associated with the esophageal mucosa.37,39,41,43–47,49,50,59,68However, metagenomic data alone cannot determine whether an organism is alive or whether only DNA traces are present. Moreover, in culture-independent approaches, sig-nificant biases could be introduced during the sample-processing steps, from the DNA extraction to the amplification and sequencing of 16S rRNA genes, which can be responsible for artifactual diver-sity in the microbiota composition.81 Fourth, the study design must be carefully considered in terms of patient characteristics, inclusion and exclusion criteria, and sampling criteria, which can be critical for the comparative analysis of results. The pres-ence or abspres-ence of GERD symptoms, for exam-ple, has been indicated among the causes respon-sible for data discrepancies.46 Moreover, exclusion criteria were highly heterogeneous or not always clearly specified, particularly concerning the use
of some drugs that can influence the composi-tion of the microbiota (e.g., antibiotics, PPIs, non-steroidal anti-inflammatory drugs, probiotics)82–84 (Tables 1–3). Another variable aspect related to the use of drugs is the time between the last drug admin-istration and endoscopy considered at risk for alter-ations in esophageal microbiota (from a minimum of 2 to a maximum of 8 weeks), but in fact no information is currently available on how long the resident microbiota could take to completely restore its original composition and structure (resilience)10 within the esophagus. Another source of variability may be the sampling methodology, which can differ in terms of the type of sample (biopsy, brush, aspi-rate) (Tables 1–3). Most studies have investigated the esophageal microbiota using tissue biopsies, which would be the most suitable specimen, considering the presence of a mucosal adherent microbiota at this body site. However, sampling by brushes or aspirates was adopted by some authors, and a less invasive sampling method has been recently pro-posed and used as an alternative to esophageal biopsies.39,50 However, despite the choice of the sampling method, esophageal biopsies are always necessary for a direct histological examination of the diseased tissues in order to avoid misclassification and require additional invasive procedures. More-over, taking esophageal samples is often associated with the risk of contamination by oropharyngeal or gastric secretions.
In this scenario, the following aspects would seem important for future investigations: (1) enroll-ment of larger cohorts of homogeneous categories of subjects and inclusion of longitudinal follow-ups; (2) expanding the studies addressed to inves-tigate the transcriptomic and metabolomic aspects of the esophageal microbiota; and (3) association of microbiota investigations with immunological investigations to clarify the potential role of micro-bial components in induction/persistence of the inflammatory status that characterizes the evolu-tion of GERD to BE and EAC. In particular, it would be important to confirm and better clarify the potential role played by dysbiosis of the esophageal microbiota or of specific pathogens in the multistep evolution model of GERD toward EAC, which still remains unclear.
Conflicts of interest
The authors declare no conflicts of interest.
References
1. Costello, E.K., C.L. Lauber, M. Hamady, et al. 2009. Bacterial community variation in human body habitats across space and time. Science 326: 1694–1697.
2. Huttenhower, C., D. Gevers, R. Knight, et al. 2012. Structure, function and diversity of the healthy human microbiome.
Nature 486: 207–214.
3. Ursell, L.K., J.L. Metcalf, L.W. Parfrey, et al. 2012. Defining the human microbiome. Nutr. Rev. 70: S38–S44.
4. Kuczynski, J., C.L. Lauber, W.A. Walters, et al. 2011. Exper-imental and analytical tools for studying the human micro-biome. Nat. Rev. Genet. 13: 47–58.
5. Morgan, X.C. & C. Huttenhower. 2014. Meta’omic analytic techniques for studying the intestinal microbiome.
Gas-troenterology 146: 1437–1448.
6. Walker, A.W., S.H. Duncan, P. Louis, et al. 2014. Phylogeny, culturing, and metagenomics of the human gut microbiota.
Trends Microbiol. 22: 267–274.
7. Kau, A.L., P.P. Ahern, N.W. Griffin, et al. 2011. Human nutri-tion, the gut microbiome and the immune system. Nature
474: 327–336.
8. Donia, M.S. & M.A. Fischbach. 2015. Human microbiota. Small molecules from the human microbiota. Science 349: 1254766.
9. Grice, E.A. & J.A. Segre. 2011. The skin microbiome. Nat.
Rev. Microbiol. 9: 244–253.
10. Lozupone, C.A., J.I. Stombaugh, J.I. Gordon, et al. 2012. Diversity, stability and resilience of the human gut micro-biota. Nature 489: 220–230.
11. Frank, D.N., W. Zhu, R.B. Sartor, et al. 2011. Investigating the biological and clinical significance of human dysbioses.
Trends Microbiol. 19: 427–434.
12. Seekatz, A.M. & V.B. Young. 2014. Clostridium difficile and the microbiota. J. Clin. Invest. 124: 4182–4189.
13. Theriot, C.M. & V.B. Young. 2015. Interactions between the gastrointestinal microbiome and Clostridium difficile. Annu.
Rev. Microbiol. 69: 445–461.
14. Johanesen, P., K. Mackin, M. Hutton, et al. 2015. Dis-ruption of the gut microbiome: Clostridium difficile infec-tion and the threat of antibiotic resistance. Genes 6: 1347–1360.
15. Martin, J.S.H., T.M. Monaghan & M.H. Wilcox. 2016.
Clostridium difficile infection: epidemiology, diagnosis
and understanding transmission. Nat. Rev. Gastroenterol.
Hepatol. 13: 206–216.
16. van Nood, E., A. Vrieze, M. Nieuwdorp, et al. 2013. Duodenal infusion of donor feces for recurrent Clostridium difficile. N.
Engl. J. Med. 368: 407–415.
17. Allen, H.K., J. Trachsel, T. Looft, et al. 2014. Finding alter-natives to antibiotics. Ann. N.Y. Acad. Sci. 1323: 91–100. 18. Gupta, S., E. Allen-Vercoe & E.O. Petrof. 2016. Fecal
micro-biota transplantation: in perspective. Therap. Adv.
Gastroen-terol. 9: 229–239.
19. Kostic, A.D., R.J. Xavier & D. Gevers. 2014. The microbiome in inflammatory bowel disease: current status and the future ahead. Gastroenterology 146: 1489–1499.
20. Gagni`ere, J. 2016. Gut microbiota imbalance and colorectal cancer. World J. Gastroenterol. 22: 501.
21. Ley, R.E., P.J. Turnbaugh, S. Klein, et al. 2006. Microbial ecol-ogy: human gut microbes associated with obesity. Nature
444: 1022–1023.
22. He, C., Y. Shan & W. Song. 2015. Targeting gut microbiota as a possible therapy for diabetes. Nutr. Res. 35: 361–367. 23. Arslan, N. 2014. Obesity, fatty liver disease and intestinal
microbiota. World J. Gastroenterol. 20: 16452.
24. van de Wijgert, J.H.H.M., H. Borgdorff, R. Verhelst, et al. 2014. The vaginal microbiota: what have we learned after a decade of molecular characterization? PLoS One 9: e105998. 25. Zeeuwen, P.L.J.M., M. Kleerebezem, H.M. Timmerman,
et al. 2013. Microbiome and skin diseases. Curr. Opin. Allergy Clin. Immunol. 13: 514–520.
26. Yang, L., F. Francois & Z. Pei. 2012. Molecular pathways: pathogenesis and clinical implications of microbiome alter-ation in esophagitis and Barrett esophagus. Clin. Cancer Res.
18: 2138–2144.
27. Wang, Z.-K. 2013. Upper gastrointestinal microbiota and digestive diseases. World J. Gastroenterol. 19: 1541. 28. Baghdadi, J., N. Chaudhary, Z. Pei, et al. 2014. Microbiome,
innate immunity, and esophageal adenocarcinoma. Clin.
Lab. Med. 34: 721–732.
29. Yang, L., N. Chaudhary, J. Baghdadi, et al. 2014. Microbiome in reflux disorders and esophageal adenocarcinoma. Cancer
J. 20: 207–210.
30. Walker, M.M. & N.J. Talley. 2014. Review article: bacteria and pathogenesis of disease in the upper gastrointestinal tract— beyond the era of Helicobacter pylori. Aliment. Pharmacol.
Ther. 39: 767–779.
31. Neto, A.G., A. Whitaker & Z. Pei. 2016. Microbiome and potential targets for chemoprevention of esophageal adeno-carcinoma. Semin. Oncol. 43: 86–96.
32. O’Hara, A.M. & F. Shanahan. 2006. The gut flora as a for-gotten organ. EMBO Rep. 7: 688–693.
33. Jandhyala, S.M. 2015. Role of the normal gut microbiota.
World J. Gastroenterol. 21: 8787–8803.
34. Lau, W.F., J. Wong, K.H Lam, et al. 1981. Oesophageal micro-bial flora in carcinoma of the oesophagus. Aust. N.Z. Surg.
51: 52–55.
35. Finlay, I.G., P.A. Wright, T. Menzies, et al. 1982. Microbial flora in carcinoma of oesophagus. Thorax 37: 181–184. 36. Mannell, A., M. Plant & J. Frolich. 1983. The microflora of
the oesophagus. Ann. R. Coll. Surg. Engl. 65: 152–154. 37. Pei, Z., E.J. Bini, L. Yang, et al. 2004. Bacterial biota in the
human distal esophagus. Proc. Natl. Acad. Sci. U.S.A. 101: 4250–4255.
38. Gagliardi, D., S. Makihara, P.R. Corsi, et al. 1998. Microbial flora of the normal esophagus. Dis. Esophagus 11: 248–250. 39. Fillon, S.A., J.K. Harris, B.D. Wagner, et al. 2012. Novel device to sample the esophageal microbiome—the esophageal string test. PLoS One 7: e42938.
40. Norder Grusell, E., G. Dahl´en, M. Ruth, et al. 2013. Bacterial flora of the human oral cavity, and the upper and lower esophagus. Dis. Esophagus 26: 84–90.
41. Pei, Z. 2005. Bacterial biota in reflux esophagitis and Barrett’s esophagus. World J. Gastroenterol. 11: 7277–7283. 42. Macfarlane, S., E. Furrie, G.T. Macfarlane, et al. 2007.
Microbial colonization of the upper gastrointestinal tract
in patients with Barrett’s esophagus. Clin. Infect. Dis. 45: 29–38.
43. Yang, L., X. Lu, C.W. Nossa, et al. 2009. Inflammation and intestinal metaplasia of the distal esophagus are associated with alterations in the microbiome. Gastroenterology 137: 588–597.
44. Liu, N., T. Ando, K. Ishiguro, et al. 2013. Characterization of bacterial biota in the distal esophagus of Japanese patients with reflux esophagitis and Barrett’s esophagus. BMC Infect.
Dis. 13: 130.
45. Blackett, K.L., S.S. Siddhi, S. Cleary, et al. 2013. Oesophageal bacterial biofilm changes in gastro-oesophageal reflux dis-ease, Barrett’s and oesophageal carcinoma: association or causality? Aliment. Pharmacol. Ther. 37: 1084–1092. 46. Amir, I., F.M. Konikoff, M. Oppenheim, et al. 2014. Gastric
microbiota is altered in oesophagitis and Barrett’s oesopha-gus and further modified by proton pump inhibitors.
Envi-ron. Microbiol. 16: 2905–2914.
47. Zaidi, A.H., L.A. Kelly, R.E. Kreft, et al. 2016. Associations of microbiota and toll-like receptor signaling pathway in esophageal adenocarcinoma. BMC Cancer 16: 225. 48. Pajecki, D. 2002. Megaesophagus microbiota: a
qualita-tive and quantitaqualita-tive analysis. J. Gastrointest. Surg. 6: 723– 729.
49. Benitez, A.J., C. Hoffmann, A.B. Muir, et al. 2015. Infla-mmation-associated microbiota in pediatric eosinophilic esophagitis. Microbiome 3: 23.
50. Harris, J.K., R. Fang, B.D. Wagner, et al. 2015. Esophageal microbiome in eosinophilic esophagitis. PLoS One 10: e0128346.
51. Wade, W.G. 2013. The oral microbiome in health and dis-ease. Pharmacol. Res. 69: 137–143.
52. Dewhirst, F.E., T. Chen, J. Izard, et al. 2010. The human oral microbiome. J. Bacteriol. 192: 5002–5017.
53. Wild, C.P. & L.J. Hardie. 2003. Reflux, Barrett’s oesophagus and adenocarcinoma: burning questions. Nat. Rev. Cancer
3: 676–684.
54. El-Serag, H.B., S. Sweet, C.C. Winchester, et al. 2014. Update on the epidemiology of gastro-oesophageal reflux disease: a systematic review. Gut 63: 871–880.
55. Yandrapu, H., M. Marcinkiewicz, I. Sarosiek, et al. 2015. Role of saliva in esophageal defense: implications in patients with nonerosive reflux disease. Am. J. Med. Sci. 349: 385–391. 56. Champion, G., J.E. Richter, M.F. Vaezi, et al. 1994.
Duo-denogastroesophageal reflux: relationship to pH and impor-tance in Barrett’s esophagus. Gastroenterology 107: 747– 754.
57. Taddei, A., V. Fabbroni, A. Pini, et al. 2013. Cyclooxygenase-2 and inflammation mediators have a crucial role in reflux-related esophageal histological changes and Barrett’s esoph-agus. Dig. Dis. Sci. 59: 949–957.
58. Osias, G.L., M.Q. Bromer, R.M. Thomas, et al. 2004. Esophageal bacteria and Barrett’s esophagus: a preliminary report. Dig. Dis. Sci. 49: 228–236.
59. Gall, A., J. Fero, C. McCoy, et al. 2015. Bacterial composi-tion of the human upper gastrointestinal tract microbiome is dynamic and associated with genomic instability in a Barrett’s esophagus cohort. PLoS One 10: e0129055.
60. Sheflin, A.M., A.K. Whitney & T.L. Weir. 2014. Cancer-promoting effects of microbial dysbiosis. Curr. Oncol. Rep.
16: 406.
61. Schwabe, R.F. & C. Jobin. 2013. The microbiome and cancer.
Nat. Rev. Cancer 13: 800–812.
62. Vogtmann, E. & J.J. Goedert. 2016. Epidemiologic studies of the human microbiome and cancer. Br. J. Cancer 114: 237–242.
63. Kusaka, G., K. Uno, K. Lijima, et al. 2016. Role of nitric oxide in the pathogenesis of Barrett’s-associated carcinogenesis.
World J. Gastrointest. Pathophysiol. 7: 131.
64. Smith, J.L. & D.O. Bayles. 2008. The contribution of cyto-lethal distending toxin to bacterial pathogenesis. Crit. Rev.
Microbiol. 32: 227–248.
65. Guerra, L., R. Guidi & T. Frisan. 2011. Do bacterial genotox-ins contribute to chronic inflammation, genomic genotox-instability and tumor progression? FEBS J. 278: 4577–4588.
66. Carbonero, F., A.C. Benefiel, A.H. Alizadeh-Ghamsari, et al. 2012. Microbial pathways in colonic sulfur metabolism and links with health and disease. Front. Physiol. 3: 448. 67. Wang, X., Y. Yang, D.R. Moore, et al. 2012.
4-Hydroxy-2-nonenal mediates genotoxicity and bystander effects caused by Enterococcus faecalis–infected macrophages.
Gastroen-terology 142: 543–551.
68. Narikiyo, M., C. Tanabe, Y. Yamada, et al. 2004. Frequent and preferential infection of Treponema denticola,
Strepto-coccus mitis, and StreptoStrepto-coccus anginosus in esophageal
can-cers. Cancer Sci. 95: 569–574.
69. Kaakoush, N.O., N. Casta˜no-Rodr´ıguez, S.M. Man, et al. 2015. Is Campylobacter to esophageal adenocarcinoma as
Helicobacter is to gastric adenocarcinoma? Trends Microbiol. 23: 455–462.
70. Man, S.M. 2011. The clinical importance of emerging
Campylobacter species. Nat. Rev. Gastroenterol. Hepatol. 8:
669–685.
71. Kaakoush, N.O., H.M. Mitchell & S.M. Man. 2014. Role of emerging Campylobacter species in inflammatory bowel diseases. Inflamm. Bowel Dis. 20: 2189–2197.
72. Tjalsma, H., A. Boleij, J.R. Marchesi, et al. 2012. A bacterial driver–passenger model for colorectal cancer: beyond the usual suspects. Nat. Rev. Microbiol. 10: 575–582.
73. Yu, G., M.H. Gail, J. Shi, et al. 2014. Association between upper digestive tract microbiota and cancer-predisposing states in the esophagus and stomach. Cancer Epidemiol.
Biomarkers Prev. 23: 735–741.
74. Nasrollahzadeh, D., R. Malekzadeh, A. Ploner, et al. 2015. Variations of gastric corpus microbiota are associated with early esophageal squamous cell carcinoma and squamous dysplasia. Sci. Rep. 5: 8820.
75. Gao, S., S. Li, Z. Ma, et al. 2016. Presence of Porphyromonas
gingivalis in esophagus and its association with the
clini-copathological characteristics and survival in patients with esophageal cancer. Infect. Agent Cancer 11: 3.
76. D’Alessandro, A. 2015. Eosinophilic esophagitis: from pathophysiology to treatment. World J. Gastrointest.
Patho-physiol. 6: 150–158.
77. Kong, J., D.B. Stairs & J.P. Lynch. 2010. Modelling Barrett’s oesophagus. Biochem. Soc. Trans. 38: 321–326.
78. Kalabis, J., G.S. Wong, M.E. Vega, et al. 2012. Isolation and characterization of mouse and human esophageal epithelial cells in 3D organotypic culture. Nat. Protoc. 7: 235–246. 79. Kosoff, R.E., K.L. Gardiner, L.M.F. Merlo, et al. 2012.
Devel-opment and characterization of an organotypic model of Barrett’s esophagus. J. Cell. Physiol. 227: 2654–2659. 80. Kapoor, H., K.R. Lohani, T.H. Lee, et al. 2015. Animal models
of Barrett’s esophagus and esophageal adenocarcinoma— past, present, and future. Clin. Transl. Sci. 8: 841–847. 81. Clooney, A.G., F. Fouhy, R.D. Sleator, et al. 2016.
Com-paring apples and oranges?: next generation sequencing and its impact on microbiome analysis. PLoS One 11: e0148028.
82. Imhann, F., M.J. Bonder, A. Vich Vila, et al. 2015. Proton pump inhibitors affect the gut microbiome. Gut 65: 740– 748.
83. Kohata, Y., K. Nakahara, T. Tanigawa, et al. 2015. Rebamipide alters the esophageal microbiome and reduces the incidence of Barrett’s esophagus in a rat model. Dig. Dis. Sci. 60: 2654– 2661.
84. Rogers, M.A.M. & D.M. Aronoff. 2016. The influence of non-steroidal anti-inflammatory drugs on the gut microbiome.
